Inference in inhomogeneous hidden Markov models with application to ion channel data

by Manuel Diehn
Doctoral thesis
Date of Examination:2017-11-01
Date of issue:2017-12-18
Advisor:Prof. Dr. Axel Munk
Referee:Prof. Dr. Axel Munk
Referee:Prof. Dr. Daniel J. Rudolf
Persistent Address: http://dx.doi.org/10.53846/goediss-6643

 

 

Files in this item

Name:Diss.pdf
Size:14.6Mb
Format:PDF

The following license files are associated with this item:

Abstract

English

Ion channel recordings under a changing environment are hardly analyzed and are the main cause for the new model class we introduce. This thesis mainly concerns hidden Markov models with a homogeneous hidden Markov chain and an inhomogeneous observation law, varying in time, but converging to a distribution. The main contribution of this thesis concerns the asymptotic behavior of a quasi-maximum likelihood estimator. In particular, strong consistency and asymptotic normality of this estimator are proven.
Keywords: Hidden Markov Models; Inhomogeneous; Strong Consistency
 
Statistik